FINPLAN: A Prototype Expert System for Personal Financial Planning.

摘要

FINPLAN is a preliminary prototype program performing personal financial planning services. It is implemented primarily as a production system using expert system concepts and runs on an IBM-compatible personal computer with 128K bytes of memory. Two principles tasks of financial planning are addressed: determining a sufficient amount of life insurane coverage and providing recommendations regarding an investment portfolio for a given client-user. The determination of life insurance needs represents a class of numerically oriented problems, which are common in the domain of personal financial planning. FINPLAN offers a solution to this class of problems by integrating numerical algorithms with production rules. Recommendations regarding the client's investment portfolio are based primarily on inference procedures and other expert system concepts. FINPLAN selects appropriate investment classes by matching the characteristics of each investmnt class with the client's own temperature and goals. The thesis gives a general description of rule-based expert systems, discusses selected design features considered in the development of FINPLAN, and suggests potential improvements that may lead to a future, more comprehensive and usable financial planning system.